Post on 01-Jan-2016
Status of Fast Pulser and Kicker Work at UIUC and Cornell
Robert Meller
Cornell University
Laboratory for Elementary-Particle Physics
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 2
Contributing Personnel :
• LEPP– G. Dugan, R. Meller, M. Palmer, D. Rubin
• UIUC– G. Gollin, M. Davidsaver, J. Calvey
• FNAL– H. Edwards, R. Fliller, J. Santucci
Overview:• Baseline ILC Damping Ring Configuration
– 2820 bunches per train with a train repetition rate of 5 Hz– Bunch spacing 3.08 ns for 6.7 km baseline ring recommendation– Fast stripline kickers
• Represent the baseline technology recommendation• Closely approach the required specifications but technology needs further validation
• Further R&D Requirements:– A bunch structure in the main linac utilizing more bunches and smaller bunch charge
has been proposed: 5000-6000 bunches per train with train repetition rate of 5 Hz– Requires higher duty cycle system
• Present efforts intended to provide input for the Reference Design Report
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 4
We chose to evaluate a new technology which has
recently been commercialized.
The Fast Ionization Dynistor (FID) is a solid-state
switch with extremely fast switch-on time and high
peak power capability.
• How long will this thing last in continuous service?
• Will the pulse repeatability be adequate?
• Will there be significant crosstalk to adjacent bunch locations?
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 5
FID Technology Catalog
Pulser: FPG2-3000-MC2
FPG1-3000 FPB3-3000 FPG10-3000
Output Impedance [Ohm] 50 100 100 100
Maximum Output per channel [kV] +/- 1 1 3 10
Number of Channels 2 1 1 1
Rise Time, 10-90% of amplitude [ns] 0.6-0.7 0.6-0.7 0.6-0.7 0.6-0.7
Pulse Duration at 90% of maximum 2-2.5 2.5-3 2.5-3 2.5-3
Fall Time 90-10% of amplitude [ns] 1-1.5 1-1.5 1-1.5 1.2-1.7
Maximum PRF in burst mode [MHz] 3 3 3 3
Maximum PRF in continuous mode [kHz]
15 15 15 15
Amplitude Stability, burst mode [%] 0.5-0.7
Pre-pulse, after-pulse [%] 1.5
Time Jitter relative to trigger [ps] 20
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 6
The unit that we obtained uses a step-recovery diode
as the pulse forming device, which is driven by a
resonant network pumped by a solid state switch.
• However, it does not contain an FID. It uses a MOSFET as the pumping device.– Our original goal of evaluating a new technology remains
untouched.– You just have to ask the right questions.
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 7
A0 Kicker, Looking Upstream
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 8
A0 Diagnostic Line
0.68 m 0.90 m 1.39 m
A0 Photoinjector Kicker Diagnostics Line
BPM BPM BPM BPM
BPM
Kicker
Corrector
SpectrometerMagnet
Beam Dump
1.46 m
0.79 m
Not all beam line elements shown
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 9
Version 1: Test Results• Present focus is on obtaining a suitable high voltage pulser
– Pulser width: wmax= 2tbunch-2tstripline ~ 4 ns, for tbunch~ 3 ns and tstripline ~ 1 ns– Have acquired a FID Technology F5201 Pulser for evaluation:
• Dual channel: +/-1 kV with 0.5% - 0.7% typical amplitude stability• 0.7 ns rise time / 2.0 ns top of pulse / 1.2 ns fall time• 3 MHz max. burst rate with 15 kHz max. average rate with <20 ps timing jitter
– Up to 10 kV devices available with similar specifications• Initial tests at A0 Photoinjector
– Stripline kicker provided by FNAL
– DAQ system provided by UIUC collaborators (G. Gollin, et al)
– Figure shows both polarity pulses as observed after kicker with 46 dB attenuation (200V/div)
– Failure in power and timing circuits during beam test
– Unit was repaired and failed again during a bench test
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 10
Timing diagnostics
• Pulser waveforms with scope externally triggered from timing support system– Note additional dispersion of
traces in time domain.
– Better timing support may be needed.
– Traces with pulser off show backward-coupled signals from passing charge bunch.
– These beam induced signals verify that the pulse overlaps the beam passage in the kicker.
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 11
Beam-Induced Signals
Backward coupled beam signals from 2.3 nC bunch,
200V/div.
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 12
Beam-Induced Signals
Forward coupled beam signals from 2.3 nC bunch,
200v/div. Magnitude comparable to backward pulse.
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 13
Beam-Induced Signals
Forward coupled beam signals from 5.2 nC bunch,
100V/div.
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 14
Version 2 of FID Pulser
Replacement pulser• Function similar to
version 1– Trailing tail somewhat
longer
• Stability– Treat full vertical
width as +/- 3 band
– Suggests ~0.7% amplitude stability, consistent with specification
– Scope is self-triggered
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 15
• Pulser waveforms triggered from A0 timing system– Oscilloscope is externally triggered
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 16
Pulser internal timing
• First pulse of self-generated train– Oscilloscope is internally triggered on pulse leading edge
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 17
Pulser internal timing
• Second pulse of self-generated train– Oscilloscope is internally triggered on first pulse
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 18
Crosstalk to cold bunches
• This type of circuit will impact preceding bunches
The 30ns charging cycle of the device couples to the output at the level of 1%
of the primary pulse amplitude (20V/div).
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 19
Time Scan of Kick Angle
• Full width: ~8.5 ns
8.5 ns
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 20
Kick Stability (Points not all sampled contiguously)Baseline Stability Sensitivity: 0.36%
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
8 10 12 14 16 18 20 22
Delay (ns)
Kic
k S
igm
a (m
rad
)Effective Kick Error
• Ideally want to verify stability at 0.1% level• Establish baseline stability (beam and DAQ) with straight through tracks at A0• Top of peak observations have about 2.5-3 times the variation of the baseline
– Timing Stability (consistent with 100ps RMS noise @ 4mrad/ns slope)– Pulser stability (potentially consistent with scope measurements and pulser specifications)
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 21
• Extrapolate A0 kick measurements to a hypothetical ILC damping ring kicker:
Simulated Effective KickExpected kick from the FPG2-3000-MC2 pulser based on the measured outputs. The black curve represents the 2.1 ns A0stripline kicker while the red curve is for a 1 ns ILC stripline kicker
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 22
Preliminary Conclusions
• Pulse Width– Full width ~8.5ns– Note that A0 kicker is ~2 ns long– With a 1 ns kicker, full width around 6.5 ns
• ILC requirement is 6.2 ns
• Pulse Stability– Appears to be near pulser specification– With a long enough kicker, the beam sees the entire
integrated pulse, and an effective flat-top occurs– With a shorter kicker, effective kick may become a
stronger function of pulse shape and timing
Sept 26, 2006 ILC Damping Rings R&D Workshop at Cornell 23
A0 Planning
• Potential Improvements– Still potentially sensitive to machine stability issues
– For stability measurements (at the 0.1% level) would like better resolution
• BPM signal processing
• BPM spacing (moment arm for angular detection)
• Better stability timing support for pulser triggering
– For higher voltage pulsers• Corrector magnet needs to be mounted around kicker vacuum chamber to
avoid scraping on limited aperture
• Also minimizes energy corrections in measurement and systematic errors from downstream BPMs
– Some additional attention to the DAQ• DAQ software needs to be able to scan
• Review of low level BPM control, such as autoranging